Performance optimization of simple exponential smoothing forecast model

<h2>Abstract</h2> In time series forecasting, the simple exponential smoothing technique allows one to forecast the next state based on serial-time past data. This model automatically assigns a weight to each instance with an exponential function, using a smoothing coefficient <mml:math><mml:mi>α</mml:mi></mml:math>, in the domain of <mml:math><mml:mrow><mml:mi>R</mml:mi></mml:mrow><mml:mo>∈</mml:mo><mml:mo>\[</mml:mo><mml:mn>0</mml:mn><mml:mo>,</mml:mo><mml:mn>1</mml:mn><mml:mo>\]</mml:mo></mml:math>. The assignment of an appropriate value of <mml:math><mml:mi>α</mml:mi></mml:math> is not a trivial task, as a transfinite defines it, and generically, a coefficient is empirically selected to find a balance between computational complexity and model error. This scientific article develops a mathematical optimization process to improve the performance of a forecasting model in time series with simple exponential smoothing using traditional optimization techniques and models derived from Artificial Intelligence. Five optimization techniques and two trained machine learning algorithms were used to find an optimized <mml:math><mml:mi>α</mml:mi></mml:math>. The results were analyzed with 20 datasets of different natures and sizes, where the model error was analyzed using four metrics that were minimized. For each optimization technique, the convergence time of the algorithm was analyzed to assess the feasibility of implementation in a production environment. Regarding running time, the computational cost is 0.01 s for machine learning algorithms on datasets of any size. In contrast, a traditional optimization algorithm converged in less than a second, even for datasets with more than 100,000 instances.